Infrared Spectra of High Polymers Part IX. Polyethylene ...

JOURNAL OF iVOLECULAR SI’ECTROS(‘OPY 3, 554-574 (1956)

Infrared Spectra of High Polymers

Part IX. Polyethylene Terephthalate

C. I-. LIANC:* AND S. KRIMM

Harrison M. Randafl Laboratory of Physics, University of Michigan, Ann Arbor, Michigan

Infrared spectra of polyethylene terephthalate and three of its deuterated analogs have been obtained in the region of 70 to 3600 cm-i, polarized spect,ra

having been obtained down to 330 cm-l. On the basis of analysis of the predicted spectrum, and with the aid of data from benzene and substituted ben-

zenes, a complete assignment of the spectrum has been made. This assignment suggests a modification of the proposed chain structure of this polymer. It also leads to the conclusion that the changes in the spectrum which occur upon crystallization of the polymer are not associated with different isomeric forms of the -OCH2CH20- portion of the chain, but rather with changes in the symmetry and resonance characteristics of the substituted benzenoid ring framework.

INTRODUCTION

The infrared spectrum of polyethylene terephthalate, (-OCH#ZH200CC6 .H,CO-). , has been the subject of fairly extensive investigation. Early efforts (1) were concerned wit,h obtaining the polarized spectrum in the region above 700 cm-L, and with the interpretation (2) of some of its salient features. Exten- sion of the spectral region down to about 100 cm-’ was undertaken (3) in order to permit a more complete analysis. The first attempt at making an assignment of bands in the infrared spectrum was due to Miller and Willis (d), who inves- tigated the region of 700 to 1800 cm-’ with polarized radiation and oriented samples. The 2600- to 3600-cm-’ region was subsequently studied (5). While a deeper understanding of the spectrum was achieved, a complete and satisfactory set of assignments did not emerge from this work. Subsequent efforts were con- centrated mainly on elucidating the changes occurring in the spectrum as a result of crystallization. Cobbs and Burton (6) presented spectra indicating pro- nounced enhancement of absorption at 1340 and 972 cm-’ upon crystallization of the polymer, as well as smaller changes in some of the other bands. This was

* Present address: Research and Development Division, American Viscose Corporation, Marcus Hook, Pennsylvania.

554

INFRARED SPECTRA OF HIGH POLYMERS 555

confirmed by Miller and Willis (7), who also pointed out that some bands are to be specifically associated with the amorphous phase. The elucidation of the structure of crystalline polyethylene terephthalate by Daubeny et al. (8) pro- vided a means for understanding a possible basis for these changes. Ward (9, 10) suggested that they arise as a result of rotational isomerism in the -OCH2 *CH,O- portion of the chain, bands at 1470, 1340, 975, and 850 cm-l being attributable to the trans configuration of the CH2 groups in the crystalline regions, while bands at 1415, 1370, 1045, and 900 cm-’ are to be associated with the gauche configuration of the CHr groups which is presumably to be found in the amorphous regions. This interpretation appeared to receive confirmation from work on linear and cyclic oligomers of polyethylene terephthalate (10, 11). Although end group assignments have been made (1.2)) and preliminary report(s on the assignments of several of the polymer modes have appeared (IS, 14), the first attempt at a detailed assignment in terms of some of the normal modes of the molecule was due to Tobin (15). Despite this thorough study, many as- pects of the interpretation of the spectrum remained uncertain.

That ambiguities of interpretation and assignment still exist is not surprising, since the polyethylene terephthalate spectrum is one of the most complex yet submitted to a detailed analysis. The simplifications of a factor group analysis (16-18) are not very helpful because of the large number of atoms in the unit cell. In order to assist in making assignments we have, in addition to obtaining far infrared and polarization data, obtained the spectra of various deuterated species of polyethylene terephthalate. A preliminary report on the latter has already appeared (19). In thin paper we wish to examine this and other data in an effort to make a complete assignment of the spectrum.

EXPERIMENTAL RESULTS

The techniques used in the present work have been described in a previous paper (17). The film xamples of normal and deuterated polyethylene terephthalate were supplied to us by E. I. du Pont de Nemours and Company. Orienta- tion was introduced by stretching or rolling. The spectra of the various polymers are shown in Figs. 1 to 4. The numbers adjacent to the curves represent specimen thickness in inches. In the polarized spectra between 330 and 3600 cm-‘, the solid curve represents absorption wit’h the electric vector perpendicular to the orientation direction and the broken curve that with the electric vector parallel to this direction. Bands are designated u or ?r according as the maximum absorption occurs with perpendicular or parallel polarized light, respectively.

The frequency, relat,ive intensity, polarization, and crystalline or amorphous origin (based on the change in intensity of the band upon crystallization) of the bands in polyethylene terephthalate are listed in Table I. The assignments given in Table I will be discussed in more detail below. In Tables II to IV are

556 LIANC; AND IiltIiVM

3600 3200 2600 2400 2000 I600 1600 1400 1200 1000 600 600 400 200 100

FREQUENCY m Cmw’

FIG. 1. Infrared spectrum of polyethylene terephthalate.

100

5 so z ?Z 60

5 z 40 2 POLYETHYLENE

I- 20 8

0 \ 3600 3200 2600 2400 2000 1600 1600 1400 1200 1000 600 600 400 200 100

FREQUENCY in cm“ _

FIG. 2. Infrared spectrum of polyethylene-D* terephthalate.

100

$ 60

; 60

g zz 40

I- 20 8

0 3600 3200 2600 2400 2000 1600 1600 1400 1200 1000 600 600 400 200 100

FREIIUENCY in cm-’

FIG. 3. Infrared spectrum of polyethylene terephthalate-D4.

60

-_ _ - I I I I I I’1 I ’

30600 3200 2600 2400 2000

FREQUENCY I” cm-’

FIG. 4. Infrared spectrum of polyethylene-Da terephthalate-Dr.

TABLE I

INFRARED SPECTRUM OF POLYETHYLENE TEREPHTHALATE

Frequency, cm-l R. I. Polarization Assignment*

N95 145 250

355

383 (c)t 430 437 (c) 502 (a) 525 (a) 575 613 (a) 633 (c) 680 (a)

7271 733(

VW

W

VW

W

m w m m W

VW

VW

W

VW

Skeletal (?)

~10#3 (B&7) (?) &s (B3U) rr(C=O) 6 (COC)

Y16B (B,,) SCCCO)

rw(C=O)

~6.4 (AI) v1os(B3,:145) + YLGB(BLU:430) = 575(B*,)

%B(BI) v;gg + S(COC) = 633

VP (BP )

S VII (B,,)

796 (a) W

845 (e) 875

898 (a) 973 (c)

1020 1043 (a)

W

m

VW

m

S

W

1100 (a)

1120 (c) 1172 (a)

1245

1263 1280

1343 (c)

rr(C=O) + S(CC0) = 792 “168 + S(COC) = 813

r,(CK) Y17B @I,)

s(COC) + yw(C=O) = 885 (?)

~12 (Bz.) VIEA (Bau)

,(CC) (‘)

!

& (B,,) .(0--C) (amorphous) Y (O-C) (crystalline)

YYA (.kl)

.(6-O,

1370 (a)

S

S

W

VW

vs

VW I S

W

.,,VA,) /e’(Ad (?)

1410

1435 1455 (a?)

S

VW

m

\vs(B3,:978) + ~,a(A,:404) = 1382(~d v,(Bs,:978) + qe~(B1u:430) = 1408(Bz,)

1473 (c) W

1504 mw 1563 VW 1580 (a?) mw 1617 (a) W

1724 vs

1830 W

1955 mw 2852 VW 2890 VW

,(0-C) + S(CC0) = 1557

YEA (Al) USE @I)

Y (C=O)

v,,,A (B,,:853) + v,,,(A,:977) = 1830(B2,) VI,.,, (A,:977) + vb(Bx,:978) = 1955(&u)

557

558 LIANC: AN0 HRIiVIbI

TSBLE I-Conknued

Frequency, cm-1 R. I. Polarization Assignment*

2908 m 2970 ms 3012 (c) VW

3055 (c) W 3068 W 3082 W 3100 W

3130 VW

3440 W

3560 W

77 v.(CHd T va (CHd ?r

?r &A (&u ) cr v(C=O) + &(&n) = 3067 (?) lr %ls(&u) lr

D 2 X v(C=O) = 3448 * v(OH) (end groups)

* Y = stretching, 6 = bending, yw = wagging, -rr = rocking. t (a) = amorphous, (c) = crystalline (data from Refs. 6, 7, 16, and 19).

TABLE II

INFRARED SPECTRUM OF POLYETHYLENE-Dd TEREPHTHALATE

Frequency, cm-i

R. I. Polarization Frequency,

cm-r R. 1. Polarization

Nl90 177

315 340 370 415

I 423 464 502 528

603 633 678 728

(734) 795 825 844 875 937 984

1019 1052 1076 1106 1122 1174

VW?

W VW m m

mw

W

m W

VW

W

VW

S

w

VW

VW

m VW m S

mw ms S

S

W

x

7r

1196 1218 1275 1375

1408 1450 1504 1526 1581

1613 1724 1828 1955 2118 2190 2247

2380 2570 2890

3055 3066 3075 3100

3450 3560

VW

VW

vs

w

S

VW

mw VW mw W

vs

W

mw m

m

mw

w

w

VW

VW

VW

VW

VW

VW

VW

w

w


TABLE III

INFRARED SPECTRUM OF POLYETHYLENE TEREPHTHALATE-DI

Frequency, cm-’


cm+

297 370

437 490 503

615

640

6%

707 754 768 781

803 816 823 844

867 898 967

1032 1068 1102

R. I. Polarization

w

mw m m m W

s

VW

W

mw

vvw VW vvw mw m mw ms VW m VW VS

VW

1166 1215

1260 1300

1335 1416

1443 1551

1586 1724

1928 1918

1

2060 2045 2280 2472 2648 2890

2960 3320 3440

3560

W

vs

S

m S

s m m W

VS

W

W

W

VW

VW

VW m VW W

W

n

listed the freyuency, relative intensity, and polarization of the bands in the

various deuterated polyethylene terephthalates.

PREDICTED SPECTRUM OF POLYETHYLENE TEREPHTHALATE

1. STRUCTURE OF POLYETHYLENE TEREPHTHALATE

A necessary preliminary to the assignment of bands in the infrared spectrum is a knowledge of the number and kinds of normal modes and their expected activity. This analysis must be based on a definite molecular structure, so we

turn first to a consideration of the st,ructure of polyethylene terephthalate. As we noted earlier, the crystal structure of this polymer has been determined

from an x-ray diffraction analysis (8). The molecules are thought to be centro- symmetric, and their arrangement in the triclinic cell of the crystal is shown in Fig. 5. The molecule is nearly planar and practically fully extended. Departures from planarity arise as a result of the COO group being about 12” out of the plane of the benzene ring, and from a rotation of the CH2-CH2 bond (around the 0-CH2 bond as axis) of about 20” from the planar configuration. From the

5cfl LI,4?;G AS 1) KltIMhl

TBBLE IV INFRARED SPECTRUM OF POLYETHYLENE-D4 TEREPHTHALATE-D4

Frequency, cm-’

317 365 386 490 502 615 640 707 728 754 763 804 816 823 844 855 894 935 982

1030 1045 1076 1112 1166 1235


cm-l R. I. Polarization

ms VW W m m W s W VW mw vvw vvw m m mw m VW W m w mw VS s W VS

P

x

K

?r

fJ

D

-

?r

-

1277 1300 1335 1340 1355 1416 1551 1586 1718 1996 2118 2130 2190 2247 2280 2310 2472 2648 2800 2965 3100 3320 3440 3560

m m

VW

VW

s W

w

VS

VW

m VW m mw VW VW VW VW VW VW VW VW W W

coordinates of the atoms it is possible to compute the orientation of various por-

tions of the molecule with respect to the fiber axis. Some of these are given in

Table V. The consistency between the spectrum and this structure, and the in-

ferences from the spectrum concerning other possible structures, will be con-

sidered later.

2. NORMAL MODES UNDER THE FACTOR GROUP

On the basis of the structure of polyethylene terephthalate described above, the factor group (do), or unit cell group (21), of the space group Pi is isomor- phic to Ci . Because of the low symmetry of this group (the only element of symmetry being a center), the factor group analysis is not very helpful in inter- preting the spectrum. It merely tells us that the mutual exclusion rule applies, and that therefore half of the normal modes of the molecule will be infrared active. In order to achieve an analysis of the spectrum we must therefore turn to a consideration of the local symmetries of portions of the molecule. This ap-

INFRARED SPECTRA OF HIGH POLYMERS ,561

FIG. 5. Crystal structure of polyethylene terephthalate [after Daubeny et al. (8)].

TABLE V

ORIENTATION OF GROUPS IX POLYETHYLENE TEREPHTHALATE MOLECULE WITH RESPECT TO FIBER AXIS%

Direction within group Angle with fiber axis

H-H axis of a CH, group Twofold axis of a CH, group -CHp-CH,- bond

\ C=O bond

/ II -C-O- bond

-0-CH,- bond Para axis of benzene ring

a Calculated from structure of Daubeny et al. (8).

77” 71”

57”

76”

47” 19” 24”

562 LIAN(; ANI) HRIivI~I

proximate method for treating t’he spectra of high polymers has been discussed

(18) and satisfactorily applied (22). In the next paragraphs we will therefore

consider the spectrum as determined by the following groups:

0

-&-C H -4 and 6 4 -0CH CH O- 2 2 .

0 0

/I II 3. KORMAL MODES OF -C-CgH4-C-

The modes of this group will be considered in terms of motions in which the

0 atoms move with respect to a rigid C-CsHh-C framework, and those in

which the C=O group moves as a unit. The former modes contribute six normal

vibrations, of which (if WC assume a center of symmetry) three are infrared active. These involve essentially the stretching of t,he C=O bond, v( C=O),

and the deformation modes in which the 0 moves parallel, rw( C=O), and perpendicular, r,.(C=O), to the plane of the benzene ring. (We assume the C- C&H,-C framework to be planar, even though the structure determinat,ion (8)

indicates a slight departure from planarity.) Their polarizations in an oriented specimen are predicted (14) to be (see Table V) u, 7r, and S, respectively.

If we assume the presence of a center of symmetry, the C-C&H,-C frame-

work will have the symmetry Vh . We choose the axis through the para substitu-

ents as the ;v axis and the axis perpendicular to the plane of the ring as the x-axis.

The character t,able, including the number of normal modes under each species,

nt , and the polarization of the infrared active bands to be expected in an ori-

ented specimen (see Table V), is shown in Table VI. The normal modes of benzene have been computed in detail (W, Z,/,), and

those for para disubstituted benzenes have been estimated (25, 26). On the

TABLE VI CHARACTER TABLE FOR C-CFH4-C

1’ h E C,(z) C*(Y) Cz(r) i

1 1 1 1 1

1 1 -1 -1 1

1 -1 1 -1 1

1 -1 -1 1 1

1 1 1 1 -1

1 1 -1 -1 -1

1 -1 1 -1 -1

1 -1 -1 1 -1

1 1 1 6 - P 1 -1 -1 6 R, - dp

-1 1 -1 2 R, - dp -1 -1 1 4 R, - dp -1 -1 -1 2 - - -1 1 1 4 T, D -

1 -1 1 6 7’, 7r -

1 1 -1 6 l’, D -


B3” $J $p$ qJi$ FIG. 6. Infrared active normal vibrations of para disubstituted benzene

basis of this work, we present in Fig. 6 a probable set of normal modes for the infrared active vibrations. The unprimed modes are benzene modes. The primed ones are modes derived from benzene modes, and represent reasonable combinations of the latter which are expected to approximate more nearly the actual normal vibrations. Although the mutual exclusion rule is expected to apply, it might be noted that the strict selection rules implied by Table VI may break down in the amorphous regions of the polymer, e.g., as a result of the two C=O groups no longer being strictly related by a center of symmetry. In this case, Ilamnn active modes may appear weakly.

4. NORMAL MODES OF -OCH2CH20-

In the above structure of polyethylene terephthalate (8) the configuration of the -OCH&H20- portion of the chain is identical with that of a trans 1,2- disubstituted ethane of symmetry Crqh . Symmetry considerations show that 9 infrared active modes are to be expected (27). From the form of the normal vibrations (28) we can determine t.he approximate nature and polarization (1,G) of the infrared active modes in an oriented specimen. Using the data of Table V, these are given in Table VII.

The above analysis is incomplete in that it predicts 12 null modes (transla- tions and rotations) whereas we know that a long chain molecule has only four

.x4 LIAN(; AND KRIMRI

TSBLE VII

APPROXIMATE NATURE AND P~LARIXATION OF THE INFRARED ACTIVE MODES OF TRANS-0CH2CHZO- a

Mode

v, (CH?) v, (CH,)

6 (CHz)

yr (CH,) ~w (CH?)

rt(CH,) v(O-C)

Chain bending: in plane Chain bending: out of plane

Polarization

a Based on structure of Dauheny et al. (8).

(I?‘, 18). It will be readily seen that when we consider the motions of the

0 0 II II

-C-C6H4-C- and -OCHzCH,O-

groups with respect to each other, 8 of the above 12 modes are internal skeletal

II motions. Among these are modes such as the stretching of the -C-O bond,

II II II Y( C-O), and the deformations of the CC0 and COC angles, S(CC0) and S(COC); the others are chain bending modes. We will consider these later.

In the following paragraphs we will discuss first the assignment of the benzene

ring vibrations, showing that a consistent analysis is possible on the basis of the

orientation of the ring in the proposed crystal structure. We will then consider the assignments of the C=O group. Finally, we will discuss the assignments for

the -OCH2CH20- portion of the molecule. It will be shown that for the latter case the spectrum is not consistent. with the proposed structure, and an slterna-

tive structure will be suggest,ed. The question of the suggested (9, 10) gauche

configuration of the -OCHKH,O- group will be considered at t’hat time.

7 Y ASSIGNMENT OF-C-C&-C- MODES

In assigning the substituted benzene ring modes we shall be guided by several principles. To begin with, we start with the assignments which have been worked out in detail for benzene (24, 29) and para dideuterobenzene (24-26). As will be shown below, modes in which the substituents do not move have frequencies very close to those in benzene. In some cases use can be made of calculations


(N4’2) of the effect of the substituent on the other modes. The deuteration results are of help in identifying some of the modes, as are pertinent correlations from smaller analogous molecules (33). We also utilize the polarization measurements in making assignments, seeing no clear-cut reason why they should be held in question (15). In fact it seems that a set of assignments consistent with the polarization measurements can be achieved for the benzene ring modes, which gives confidence that the polarization results are also applicable to the other modes. Finally, it must not be overlooked that relaxation of the Vh symmetry, such as could occur in the amorphous regions, will result in the appearance of some of the normally infrared inactive Raman modes.

1. ASSIGNMENT OF Bzu MODES

We consider first the assignment of the Bzu modes, since they should be readily idemifiable by their r polarization. The five Bxu modes derive from two inactive B1, modes ( ~12 and ~13) and three infrared active El, modes ( v18 , vlg , and vzo)

of benzene. The v1aA mode should be readily identified since the substituents in the para

position do not move. It occurs as a strong band at 1037 cm-’ in benzene, and we have no difficulty in assigning to it the strong ?r band at 1020 cm-l in the polymer. It probably shifts to about 860 cm-’ on ring deuteration, in comparison with the shift to 819 cm-l in 2,s ,5,6_tetradeuterobenzene (26). The CH stretching mode, v:OA , can similarly be identified with the r band at 3055 cm-l. It is found at 3060 cm-’ in paradideuterobenzene (28). The v19A mode occurs at 1485 cm-l in benzene, and the evidence seems to indicate (23, SS, 34) that it moves to slightly higher frequencies in para disubstituted benzenes. On this basis the vlgA mode can be readily assigned to the a band at 1504 cm-‘. From its disap- pearance on ring deuteration, and the relative enhancement and slight shift of the band near 1410 cm-‘, deuteration appears to shift it to about 1415 cm-‘. This indicates that this mode contains a significant proportion of H motion, perhaps as shown in Fig. 6.

The identification of the ~12 and v;, modes is less certain. The latter, as indicated, is expected (26) to be a stretching mode of the CC bond between the benzene ring and the C=O group. Because of the planarity of the

II II -C-CsH4 - C-

framework, and the resulting resonance in this conjugated structure, we expect this CC bond to have some double-bond character. This should place the vi, mode at a higher value than if a pure CC single bond were involved (for example, in butadiene the v(CC) mode is found (55) at 1205 cm-‘). We believe that Y;, should be assigned to either of the bands at 1343 or 1410 cm-‘. Since we

566 LIANG ANI) KRIMhl

will show later that the 1410 cm-’ band is most probably a combination band, we choose the band at 1ZX~ cm-‘. I:or various re;lsons this band camlot be assigned to a CH, mode, such as has been suggested or implied by many authors

(4, 9, 10, 16, 19). At present we will just not,e that this band is nob shifted by deuteration of the CH, group (see Fig. 4). It may seem from Fig. 2 that deuter-

ation has caused this band to disappear, but as we shall presently discuss, there

is good reason to believe that it has merged with the band at 1263 cm-‘, which,

in polyethylene-D4 terephthalnte (Fig. 2) has broadened and shifted t,o 1275

cm-‘, the highest value it assumes in the four polymers. The absence of a ahift on ring deuteration is quite consistent with the nature of this mode.

Before proceeding with the other assignments, it is necessary to say a few words about the significant increase in intensity of the lSK~-cm-l band which

we have noted occurs upon crystallization of the polymer. This behavior is quite

consistent w&h the proposed assignment.. If, as we shall show later, the polymer in the amorphous regions is characterized by a loss of its center of symmetry

(as evidenced by the appearance of Raman active bands), we would expect this 0 0

II II to imply a departure from planarity of the -C-C&H, - C- group. This in turn

will be associated with a loss of resonance in the CC bond whose stretching is involved in the & mode, and t’he position and intensity of this mode would be

expect’ed to change. The increased intensity upon crystallization is then just a 0 0 II II

reflection of the increased number of -C--CcH,-C- groups with strict Bh symmetry. If this effect occurs for the & mode, we shall expect it in other modes,

especially those in which the motions significantly involve this CC bond, and

this in fact seems to be the case. It might be noted that the suggested merging of this band with that at 1263 cm-’ in polyet’hylene-D4 terephthulate is not at all

II unreasonable. We will see that t,he latter band is to be associated with Y( C-O),

II and it is quite likely that under certain conditions V( C-O) and & could be quite close to each other.

The final mode in the BZa species, v12 , is an inactive mode in benzene located

at about 1010 cm-‘. Its assignment in polyethylene terephthalate is difficult to make with certainty, but we believe that the position, intensity, polarization, deut,eration behavior, and intensity increase upon crystallization of the band at 973 cm-’ are consistent with its assignment to v12 .

2. ASSIGXMENT OF BI, MODES

The three B1, modes derive from one infrared active A, mode (~11) and two inactive Ezta modes ( v16 and ~17) of benzene. The ~16~ mode is found at 405 cm-l in benzene, and is expected in the same general region (23, 36) in para disub-


stituted benzenes. Since it should exhibit u polarization in oriented polyethylene terephthalate, the only possibilities for assignment are bands at 355 and 430 cm-‘. The former is more reasonably assigned otherwise. Choice of the latter will be shown to be consistent with certain expected combinations. We therefore assign the band at 430 cm-’ to v16B .

The two out-of-plane bending modes will be expected to give rise to moderately strong u bands in the region of approximately 600-900 cm-’ (24, 32, 33, 37). Despite its early misassignment (4), it soon became evident (13, 1.5) that the band at 730 cm-’ is to be associated with one of these modes. Tilting experi- ments on oriented specimens (1:3) also showed that the 875 cm-1 band is associated with an out-of-plane bending mode. Ring deuteration (19) confirms this assignment. From the deuterated spectra (see Figs. 1 and 3) it is clear that the 730~cm-’ band shifts to 640 cm-‘, and the 875-cm-’ band shifts to near 820 cm-l (an assignment of the shift to 707 cm-’ leaves the moderately strong u bands near 820 cm-l completely unassignable). On the basis of the larger isotope shift and greater intensity of the 73O-cn~-’ band we assign it to vll , whereas vlyB is associated with the band at 875 cnl-I.

3. ASSIGKMENT OF B3,& MODES

The five BSu modes derive from two inactive Bzu modes ( v14 and v16) and three infrared active El, modes ( v18 , v19 , and VZO) of benzene. The v2OB mode, which is found in benzene at 3080 cm-’ and is expected to change hardly at all in para

disubstituted benzenes, is readily identified by its position and polarization with the band at 3082 cnli’. The vlgB mode should be located near 1450 cm-‘, but is often not found (33). We believe that it may be associated with the weak crystallization-variable band at 1473 cm-‘. It is the only band with appropriate polarization in the expected region, and its enhancement with increasing crystallinity is consistent with the arguments set fort’h above in connect,ion with the assignment of &( &). (A possible assignment to a s(CH2) mode will be considered later.)

The assignments of two of the other three modes can be made with fair certainty. On the basis of analogy with p-xylene (23), where the band is thought to be located at 232 cnii’, we would assign the vi,, mode in the polymer to the weak band at 250 cm-‘. The &, mode should be essentially a hydrogen mode (as), and should be found near 1100 cm-l (23, 26). From the appearance of T bands in this region in polyethylene terephthalate it would seem that no assignment can be made. However, it has already been noted (IO) that this group of bands is of quite complex origin. In fact weak bands near this position are found in terephthalic acid (11 ), which would not be expected to exhibit the v(O-C) mode which is the main contributor t’o this region in the higher oligomers (where

II the intensity relative to the very strong v( C-O) mode near 1250 cm-’ remains

essentially constant) and in polyethylene terephthalnte. We therefore feel that there is good reason to believe that, Y & contributes to the band at 1100 cm-‘. Because of it,s weakness, its u polarization is not evident; the 7r polarization of

the overlapping band dominates. This assignment is further supported by the presence of a weak c band at 935 cm1 in the ring deuterated polymer (see Fig. 4) which bears the same isotopic shift ratio to t.he 1 lOO-cm-’ band as is the case

for the analogous CH bending mode Ylg.,, . The assignment of t,he remaining mode

of this species, ~~4 , is in doubt in benzene (a.$), and we have been unable t,o locate it with any assurance in our polymer spect,rum.

4. ASSIGNMENT OF RAM~N ACTIVE MOPES

The assignments discussed in the above paragraphs still leave many bands, which are not associated wit)h C=O, -OCH2CH20--, or skeletal modes, un-

accounted for. We will show in t,his and the next section that these bands are due to Raman active modes which become infrared active as a result of loss of symmetry, and to combination bands.

As we have observed before, loss of the center of symmetry, e.g., by rotations of the C=O groups out of the plane of the benzene ring, will lead to the appear-

ance in the spectrum of bands which are forbidden under Vh symmetry. It should

be possible to identify such bands by three characteristics: knowledge of their positions from Raman data, the fact that their int,ensity increases with increase

in amorphous content, and their polarization, which should be correlatable wibh the activity predicted under a lower symmetry, which would most probably be Czv (22). For example, relaxation of symmetry should give rise (33) to the ap-

pearance of bands associated with vx , which appears at 1585 cm-’ in benzene. The evidence strongly suggests that the bands at 1580 and 1617 cm-‘, which increase in intensity with increasing amorphous content, are due to the @.d(lllj

and vSB(BI) modes, respectively. (For the form of these normal modes, see Refs. 22 or 23.) These modes are actually strongly mixed with ~9 (24), which accounts for the small shift to 1551 and 1586 cm-‘, respectively on ring deuteration. Other modes which seem to fit into this category (3.9) are ~9.~ (A,) at 1172 cm-‘, vJ( B2)

at 680 cm-‘, vGR(B1) at 613 cm-‘, and possibly vgA(Al) at 525 cm-‘. Other, less

certain, possibilities are VZ’ at 1370 cm-’ [t,his is the Raman active counterpart

of & , and is found at a slightly higher frequency than the infrared active mode in para dideuterobenzene (25, 26)]; v(CC), the stretching of the H2C-CH2 bond, at 1043 cm-‘; and Y~O~(&,) at 145 cm-‘. The reason for suggesting the latter assignment is that yloB is expected to occur at a low frequency (W), and its assignment as suggested leads to the accurate reproduction of a combination band observed in benzene (23).

5. ASSIGNMENT OF COMBINATION BANDS

It is known that in benzene (23) and in substituted benzenes (38) certain characteristic combination bands of the out-of-plane CH bending vibrations


appear with significant intensity in the infrared spectrum. This was found to be true in polystyrene (dd), and we would expect a similar situation in the spectrum of polyethylene terephthalate. This will be shown to be the case, which also aids in confirming some of the assignments of the fundamentals.

In assigning combination bands, we are guided by two principles. The species

of the combination should follow from those of its components according to the character table, and Raman active components should have frequencies near

those of t’he corresponding mode in benzene when this mode is one in which

there is no motion of the para substituents. In the latter connection it is also

often possible to make use of Raman data on similar smaller molecules, such as

dimet.hyl and diethyl terephthalate (39).

In this way, for example, we can identify two characteristic combination bands

at 1830 and 1955 cm-‘. The former is a ?r band and therefore belongs to the Bzl, species. One of the characteristic combinations is ~10 + ~17 , and it will be readily seen that yloB + vlTA (849 + 984 = 1833 in benzene) gives a band in the appropriate region and of t#he correct species. If we use the value of 853 cm1 for

vlOA from diethyl terephthalate (S9), this would place ~17~ at about 977 cm-‘. The G band at 1955 cm’ must belong to the B1, or Bau species. Again it can be seen that the characteristic combination Y17A + ~5 gives a band (on the basis of

the benzene frequencies) which is in the appropriate range and of the correct

species. Using the above value for ~~7.~ we find that vg is at about 978 cm-‘, not

far from it’s value of 967 cm-’ in para dideuterobenzene (25). Using this value

for v5 and our assigned value for v16B we find t’hat another characteristic com-

bination, v5 + VM , is accurately reproduced. It might be thought that the Ill@cm-’ band is too strong to be attributed to a combination band, but in fact such combinations are quite strong in benzene (29). Finally, vg + vlGd may

be contributing t,o the 1370-cm-’ band. The other possible combinations of t,hese characteristic out-of-plane CH bending modes give bands which, other than the

apparently absent v5 -I- v]?B , overlap with other strong bands in the spectrum. In summary, it appears that the spectrum predicted for the C-_C6H,-_C

framework is satisfact’orily identifiable in detail within the spectrum of poly-

ethylene terephthalate. Since this has been based on the assumption of an ori-

entation of the benzene ring similar to that in the crystal structure, we may presume that this orientation (viz., with the para axis of the substituted ring oriented essentially parallel to the stretching direction) is substantiated by the

spectral result’s We have also seen t’hat the intensity of certain of the ring modes is sensitive to the exact symmetry of the ring, which seems reasonably to be associated with the amorphous or crystalline state of the polymer.

In the category of modes associated with the C=O group we expect to find three bands, and their identification is not too difficult. There is no question about the location of v(C=O) at 1724 cm’. The rr(C=O) mode is expected to give rise to a u band. On the basis of its assignment in ketones to a band at,

570 LIANG AND KItIMM

390 cm-r (4(I), we feel that it is probably to be assigned t,o the 355-cm’ c band in polyethylene terephthslate. The r,,.(C=O) mode is assigned to a band at 639 cm-’ in CH&OOCH, (41) and at 527 cm’ in CH.$OCHaCOCHS (40). The assignment in the polymer is not ent#irely unambiguous, but, we believe t,his mode can reasonably be assigned to the ?r band at 502 cm-‘. Its supposed enhancement in the amorphous state (15) may be related to the interaction between the C=O group and the benzene ring, which we discussed earlier.

ASSIGNMENT OF --OCH&H,O- ANI> SKELETAL MODES

In discussing these assignments, we will first try to dispose of those modes II

involving stretching and deformation vibrations of the -C-O-C- group, reserving for the subseyuent discussion the assignment of t,he CH, modes and the attendant questions of the configuration of the -OCH&H,O- portion of the chain.

1. ASSIGNMENT OF -C-O-C- MODES

Esters are known (41, 42) to give rise to two strong absorption bands asso- II

ciated with the stretching of t’he C-O and O-C bonds. These can be assigned with a fair degree of certainty to the strong ?r bands at 1263 and 1120 cm-‘, respect,ively. The smaller dichroism of the former is consistent with the less

II parallel orientation of the C-O bond (see Table V). The presence of two r bands in the llOO-cm-’ region may be an indication of the dependence of t,he ,(0--C) mode on the configuration of the neighboring CH2 groups with respect

0 II

to the CsH4-C- portion of the chain. This is hypothetical at present, but we will shortly consider some arguments which point in this direction.

Assignments have been suggested (41) for the S(CC0) and 6(COC) modes. These are consistent with the bands which we find in the 300-500 cm-’ region. The difficulty about making a reasonably certain assignment is that we do not know t,he forms of the normal vibrat’ion. It would seem, on the basis of a simple three mass model, that these deformation vibrat,ions should exhibit r polarizn- tion. However, it is well known (43) that the actual form of this vibration is a sensitive function of the masses, and in fact for certain mass distributions this mode will exhibit ?r polarization (Ref. 68, pp. 225-227). We will therefore assume that this is the case, and base our assignments on those for CHSOOCH, (41 I. This leads to the assignment of the 437~cm-’ band to S(CC0) and the 383-cm-’ band to 6( COC). Perhaps the apparent weakening of the 437-cm-’ band on ring


deuteration is associated with an interaction between this mode and one of the ring modes. We are unable to indentify with any certainty the skeletal modes, and suggest that they may be the origin of perhaps one of the weak low-frequency

II bands in the spectrum. Several combination bands involving these C-4-C modes seem to occur in the spectrum, and are indicated in Table I.

2. ASSIGNMENT OF CH2 MOUES

The assignments of the v,( CH,) and Y,(CH~) modes can be made on the basis of their well-established positions in hydrocarbons (44), where they are found at 2853 and 2925 cm-‘, respectively. If we take into account the increase in these frequencies which occurs when the CH2 group is adjacent to the ester linkage (/,5), then there is no difficulty in assigning v~(CH~) and v,(CH~) to bands at 2908 and 2970 cm-‘, respectively. What becomes immediately apparent is that their ?r polarization is inconsistent with the c polarization predicted from the crystal structure (8, 14) ( see Table V). We shall return to this point presently. There may be an additional u band in this group, since a third a band appears in CD2 polymers (see Figs. 2 and 4). The 6( CH2) mode is similarly (44) located with no difficulty at 1455 cm-’ , shifting to 1076 cm-’ on deuteration. Its s polarization is consistent (l/t) with that of the Y.(CHZ) mode. The possible presence of a g v(CH2) mode suggests that the g band at 1473 cm-’ may be a 6( CH,) mode. This is difficult to verify by deuteration because of the weakness of the band.

The assignment of the rU-(CH,) mode is of especial importance, since it has been made the basis of a postulated gauche configuration for the -OCH,CH20- group in the amorphous regions (9, 10, 19). On the basis of the existence of similar bands in the spectra of rotational isomers of ethylene dihalides (46) and in polyethylene glycol (47)) Ward (9, 10) has postulated that the 1343- and 1370- cm-l bands are associated with T~( CHZ) modes of the trans and gauche forms of the -OCH,CH20- group, respectively. We wish to point out that these bands are probably not due to CH2 groups. First, as we have already noted, the 1333- cm1 band does not shift on deuterat’ion of the CH2 groups. The spectra indicate that the same is true of the 1370-cm-’ band. Second, the Y,~(CH~) mode, which is located at about 1369 cm-’ in hydrocarbons (&), is generally very weak in polymers in comparison to the 6( CH2) mode (22, 44, 48-50). Although in some cases its intensity appears to be enhanced (51), it is never more intense than the 6(CH2) mode. On this basis it would be difficult to assign the 1343-cm-’ band to yw (CH,). Third, as we have shown (14)) in terms of the strong T v,( CH,) and v,( CH,) modes the -yw( CH2) mode should exhibit (r polarization. This also eliminates the 1343cm-’ band from consideration, leaving only the 1370-cm-’ band. The latter may be due to r,,.(CH,), but at present we think this unlikely,

considering t.hat it does not, shift on deuteration of t,he CH:! group. We would therefore conclude that rw( CEI,) is not detected in the spectrum, :utd would a(:- count. for the intensity changes of t:he 1X3-cm-’ hand not, by rot.at.ional isomerism of the CH, groups but, as indicat,ed in the earlier discussion, by the influence

0 0 II II

of resonance in the -C-C&-C- group on certain of the benzene ring modes. On the basis of the crystal structure of polyethylene tcrephthalate (8) the

-y*(CH,) mode should give rise to a u band. There is no band in the spectrum in the appropriate region which can thus be assigned. If, however, we follow the observed 7r polarization of the intense v,( CH2) and v:,( CH2) modes in expecting (14) a a rr(CH2) mode, then the assignment of t’his mode to the n band at, 815 cm-’ is a reasonable one, particularly since this band appears to shift on deutera- t’ion of t’he CH, group. The only other CH, mode, rt( CH2 ), the CH2 t’wisting mode, is generally too weak to be observed, and we have not been able to assign it in the present work.

STRUCTURE OF POLYETHYLENE TEREPHTHALATE

We wish to consider now the implications of the above result,s with respect, to the structures that have been suggested for the crystalline (8) and amorphous (9, 10) forms of polyethylene terephthnlate.

It has already been noted t,hut, whereas the spectrum is consistenC with t)he benzene ring orientation proposed for t’he crystalline structure, this is not the case for the CH? groups. Their modes exhibit dichroism in complete disagreement with that predicted, and, if the assignment of t,he 84%cm-’ crystalline hand to rr(CH2) is correct, t.his diehroism seems t,o be associated with the ery&lline phase. Since we find no reason t,o hold t.he polarization data in question, nor their correlation with the t,rnnsit’ion moment,s of the CH, group, we are led t,o suggest that the proposed chain structure may be in error with respect to the orientation of the CH2 groups. It is interesting to note that the authors t,hem- selves (8) point out that, the location of these carbon atoms was the least, reliable part of the structure detjermination. The dichroism of the CH, modes can be nc- counted for by a change in the structure which, molecular models indicate, is easily possible. This consists of an additional rotation of the CHzCHa bond about the 0-CH2 bond followed by a rotation about t’he C-O bond. This still pre- serves the center of symmetry and gives a fiber axis repeat essentially ident’ical with that observed. It should be noted that such a larger rot’ation about the 0-CH, bond is observed in polyethylene adipate (8, 52?), a fact which indicates that rotation about this bond occurs with relative ease. We feel that the interpretation of the infrared spectrum is compelling enough so that the possibility of the predominant, structure being such as that suggested here must be seriously considered.


The second point concerns the presence of a gauche configuration for the -OCH&H20- group in the amorphous regions. As will have been gathered by now, t’he changes in the spectrum upon which the above conclusion is based are not associated with CH2 modes, as claimed (9,10,53). Such prominently crystallization sensitive bands as those at 973 and 1343 cm-’ are not CH2 modes, since they are not affected by deuteration in the expected manner. While it is possible that the chains in the amorphous regions have gauche -OCH&JH20- groups, we do not think that unambiguous evidence exists in the infrared spectrum in support of this. Rather we believe that the changes that occur upon crystallization are associated with changes in the symmetry and resonance characteristics

0 0 II II

of the -C-C$H4-C- framework. In addition, in view of the discussion of the previous paragraph, it is probably more likely that structural changes in the chains in the amorphous regions are associated with rotations around the 0-CH2 bonds than with discrete rotationally isomeric states about the CH2-CH2 bond.

ACHNOWLEDGMEXTS

The authors wish to express their appreciation to E. I. du Pont de Nemours and Com- pany for supporting grants, during the tenure of which part of this work was done. We are also indebted to this company for making available t.o us samples of deuterated polyethylene terephthalate. We wish to thank I)rs. W. W. Daniels and R. E. Kitson for the oppor-

tunity of reading the manuscript of their paper prior to publication, and Dr. H. W. Wyckoff for helpful discussions on the molecular model of polyethylene terephthalate. We are grate- ful to Dr. 6. B. B. M. Sutherland for helpful discussions.

~~ECEIVED: October 7, 1958

REFERENCES

1. R. BARER, A. R. H. COLE, AND H. W. THOMPSON, Nature 163, 198 (1949). 2. J. MANN AND H. W. THOMPSON, Proc. Roy. Sot. Aall, 168 (1952). 8. S. KRIMM AND G. B. B. M. SUTHERLAND, Symposium on Molecular Structure and Spec-

troscopy, Columbus, Ohio, 1952. /t. R. G. J. MILLER AND H. A. WILLIS, Trarzs. Faraday Sot. 49, 433 (1953). 5. G. J. WESTON, Chem. and Znd. p. 604 (1954). 6. W. H. COBBS, JR., AND R. L. BURTON, J. Polymer Sci. 10, 275 (1953). 7. R. G. J. MILLER AND H. A. WILLIS, J. Polymer Sci. 19, 485 (1956). 8. R. DE P. DAUBENY, C. W. BUNN, AND C. J. BROWN, Proc. Roy. Sot. A226, 531 (1954). 9. I. M. WARD, Chern. and Znd. p. 905 (1956).

10. I. M. WARD, Ghem. and Znd. p. 1102 (1957). 11. B. SEIDEL, Z. Elecktrochem. 62, 214 (1958). 12. D. PATTERSON AND I. M. WARD, Trans. Faraday Sac. 63, 291 (1957). 15. C. Y. LIANG AND S. KRIMM, J. Chem. Phys. 27, 327 (1957,. 14. C. Y. LIANG AND S. KRIMM, J. Chem. Phys. 27, 1437 (1957). 16. M. C. TOBIN, J. Phys. Chem. 61, 1392 (1957). 16. M. C. TOBIN, J. Chem. Phys. 23, 891 (1955).

574 LIANG AND KRIMM

17. C. Y. LIANG, S. Iimww, ANII (:. 13. B. M. SUTIIERLAUD, J. (‘her,/. Phi/s. 26, 543 (1956). 18. C. Y. LIANG, J. Mol. Spectroscopy 1, 61 (1957). 19. W. W. DANIELS AND It. E. KITSON, J. Polynle, Sci. 33, 161 (195s). do. S. BHAGAVANTUnl AND T. VEXKATARAYUI)~, “Theory of Groups and its Applicstion to

Physical Problems,” Andhra University, Waltnir, India, 1951. 91. D. F. HORXIG, J. Chem. Phys. 16, 1063 (1948). 22. C. Y. LIAxG AND S. KRIMM, J. Polynler Sci. 27, 241 (1958). 28. K. S. PITZER AND D. H. SCUTT, J. Bnh. Chem. Sot. 66, 803 (1943). 24. D. H. WHIFFEN, Phil. Trans. Roy. Sot. A248, 131 (1955).

25. N. HERZFELD, J. W. H~BDEX, C. Ii. INGOLD, AND H. G. POOLE, J. Chem. Sot. p. 272 (1946).

26. C. R. BAILEY, S. C. CARSOX, R. R. CORDON, AND C. K. ISGOLD, J. Chem. Sot. p. 288 (1946).

27. S. MIZ~SHIJXA, “Structure of Molecules and Internal Rotation,” p. 33. Academic Press,

New York, 1954.

28. IX. W. F. KOHLRAUSCH, “Ramanspektren,” p. 177. Akademische Verlagsgesellschaft

Becker and Erler, Leipzig, 1943. 29. R. D. MAIR ASU D. F. HORNIG, J. Chem. Phys. 17, 1236 (19-l9).

SO. E. HERZ AND J. WAGNER, Monatsh. Chem. 76, 93 (1946). 81. E. HERZ, K. W. F. KOHLRAUSCH, AND R. VOGEL, Monatsh. Chem. 76, 200 (1946). 36. A. R. H. COI,E AND H. W. THOMPSON, Trans. PCATU~U~ Sot. 46, 103 (1950). 33. R. R. RAXULE AND D. H. WHIFFEN, “Molecular Spectroscopy-,” George Sell, ed., p.

111. Institute of Petroleum, London, 1955.

34. E. E. FERGLXON, R. L. HUDSON, J. R. NIELSEN, AND 1). C. S~XITH, J. Chem. Phys. 21,

1457 (1953). 35. c. hf. RICI!ARDS AND J. R. NIELSEN, J. Opt. SOC. A?n. 40, 438 (1950). 36. E. K. PLYLER, Disc. Faraday Sot. 9, 100 (1950). 37. T. SHInrANoucHI, Y. KAICI~TI, AND I. GAMO, J. Chem. Phys. 26, 12-45 (1956). 38. D. H. WHIFFES, Spectrochim. Acta 7, 253 (1955). 89. A. PONGRATZ AND It. SEKA, Monatsh. Chew 66, 307 (1935). 40. J. LECOMTE, Disc. Faraday Sot. 9, 125 (1950). 41. J. K. WILDISHURST, J. Mol. Spectroscopy 1, 201 (1957). 42. H. W. THOMIJS~N AND P. TORKINGTON, J. Chem. Sot. p. 640 (1945).

43. G. HERZBERO, “Infrared and Raman Spectra,” p. 174. Van Nostrand, New York, 1945. 44. S. KRIMM, C. Y. LIANG AND G. B. B. M. SUTHERLAND, J. Chem. Phys. 26, 549 (1956). 45. B. XOLIN AND R. N. JONES, Can. J. Chem. 34, 1392 (1956). 46. J. Ii. BROWN ANU N. SHEPPARD, I’rans. Faraday Sot. 48, 128 (1952). 47. W. H. T. DAVISON, J. (Them. Sot. p. 3270 (1955). 48. S. KRIMM AND C. Y. LIANG, J. Polymer Sci. 22, 95 (1956). 49. S. KRIMM, C. Y. LIANG, ANU G. B. B. M. SUTHERLAND, J. Polymer Sci. 22, 227 (1956).

60. 0. THEIMER, J. Chem. Phys. 27, 1041 (1957). 61. C. Y. LIANG AND S. KRIYM, J. Polymer Sci. 31, 513 (1958). 5%‘. C. W. BUNN, J. PoEymer Sci. 16, 323 (1955). 58. D. GRIME AND I. M. WARD, Trans. Faraday Sot. 64, 959 (1958).

Infrared Spectra of High Polymers Part IX. Polyethylene ...

Documents

Transcript of Infrared Spectra of High Polymers Part IX. Polyethylene ...